Zach Rattner - Yembo

Yembo: From Cold Calls & Rejections to Scaling an AI Startup – with Zach Rattner [400]

Yembo: From Cold Calls & Rejections to Scaling an AI Startup

Zach Rattner is the co-founder and CTO of Yembo, an AI-powered platform that enables virtual home surveys for the moving and insurance industries.

In 2015, while working as a software engineer, Zach noticed that computers were becoming better than humans at identifying objects in images.

His wife's experience working at a moving company inspired him to apply this technology to the industry, which struggled with giving accurate quotes and handling logistics due to the complexities involved in each move.

However, building an AI-powered product was no easy feat.

As introverted engineers, Zach and his co-founder Sid had to force themselves to step out of their comfort zone. They made cold calls, visited moving companies in person, and often faced rejection.

In the early days, the founders also handled sales themselves. They attended industry trade shows and conferences to generate leads and build relationships with potential customers.

Despite their efforts, the first version of Yembo's product had limitations in its AI capabilities and user interface, which led to some customer churn.

The founders realized they needed to focus on finding early adopters willing to work through initial challenges and continuously iterate based on customer feedback.

Through their determination and hard work, Yembo gradually gained traction.

Today, the company serves customers in about 30 countries, processing hundreds of hours of video daily and generating high seven-figures in annual revenue with a team of 70 people.

In this episode, you'll learn:

  • How Zach and Sid validated their idea despite facing rejection and discomfort, and what strategies they used to overcome these challenges.
  • Why setting realistic expectations about your AI-powered product is crucial for maintaining trust and preventing disappointment among customers.
  • How Yembo overcame skepticism by educating customers about advancements in technology and focusing on early adopters eager to try new solutions.
  • Why handling sales yourself initially, even as a technical founder, is essential for gathering direct feedback and refining your pitch.
  • How attending trade shows and creating engaging demos can help generate leads and build relationships with potential customers, even in underserved markets.

I hope you enjoy it!


Click to view transcript

This is a machine-generated transcript.

[00:00:00] Omer: Zach, welcome to the show.

[00:00:01] Zach: Hi, Omer. Happy to have, happy to be here. Thanks for having me.

[00:00:04] Omer: I'm so glad we finally did this. If, if the listeners knew like how long it's taken us to get this thing scheduled and, and kind of reschedule it's quite an achievement. So thank you for making the time today.

[00:00:17] Zach: For sure. If something's important, it's worth doing.

[00:00:19] Omer: Yeah, totally. Yeah. Do you have a favorite quote, something that inspires or motivates you that you can share with us?

[00:00:24] Zach: I do. I am a big big fan of the band, Radiohead. And they have a lot of good lines where if you digest what they're saying, it's not always like immediately apparent, but there's a line on the album, kid A, this one's optimistic, this one went to market.

[00:00:41] And it talks about all the hosts of things in general that can go wrong, but I I feel like as a founder, being optimistic and still going to market anyway is something that kind of gets me outta bed in the morning.

[00:00:51] Omer: Yeah, love that. So tell us about Yembo. What does the product do? Who's it for and what's the main problem you're helping to solve?

[00:00:58] Zach: Sure. Yembo is a computer vision company. We are about 70 folks across the world, about half in the us, half out, and we provide computer vision services. To moving companies and property insurance companies. So in both of these markets, if you want to get a quote for something, it's traditionally very labor intensive.

[00:01:18] Someone has to schedule some time, ring your doorbell, walk around, note down every item that's being moved, or every item in the policy if you're getting insurance. And what we provide is a computer vision workflow where. You can send your clients a link. They can record quick videos of each room in the house, and then the AI identifies what's there, pulls out key attributes for moving volume and weight are pretty key, and then we provide that information back so you can have an accurate quote with the real photos of the items being serviced in there.

[00:01:48] So our business model is to sell to the service provider. So we sell to moving companies and sell to insurance companies. But we provide the whole suite of software where they use it for their end users, but the end customer gets it for free effectively 'cause our, our client is the company that's providing the service.

[00:02:06] Omer: Great. And give us a sense of the size of the business. Where are you in terms of revenue, customers, size of team?

[00:02:13] Zach: Sure. So we have customers in about 30 countries across the world. Probably 70% of our business is in North America, but do have a sizable presence in Europe and Asia as well. We have every day we're processing a couple hundred hours of video.

[00:02:28] Those can be quick little 20, 32nd recordings of each room all the way on up to longer calls for live video chats. And in terms of revenue, we're in the high seven figures. And probably, every day we see like, on order of a couple thousand inspections being done across the whole product suite.

[00:02:48] Omer: Cool. You, you mentioned you're a team of, about, I think 70 people. What, what's the general makeup of the team? I mean, with a AI startup do you, do you kind of lean heavily towards people working on the technology or do you have a big sales team or, you know, how, how is kind of the general setup of the team?

[00:03:05] Zach: Yeah, so we are, we are We are a engineering company. I'm an engineer by training. My co-founder, the CEO also is an engineer. Our team is about half engineers. The other half we have operations, customer success, go-to market folks. We found that's really key because at least in the business world, we're not selling.

[00:03:27] Something that's necessarily fun or a game, like we want to provide a delightful user experience, but being implementation experts is really key. So that's why we have a pretty big team that understands the client workflows and we sort of become like expert management consultants almost, that we're not just trying to sell software, but we're explaining here's how you can take cutting edge 21st century technology and bring it back to these traditionally underserved by tech communities.

[00:03:54] And that's why we've ended up having kinda about 50-50 engineering versus not just nature of the business that we're in.

[00:04:00] Omer: Cool. And you know, you guys were founded in 2016 and you were an AI company before everybody was trying to be an AI company, right? So tell us about like where you came up with the idea for, for this business.

[00:04:17] Zach: Sure. So rewind the clock a bit. It's maybe 2015 or so, and there is this academic benchmark, it's called ImageNet, and if you're not familiar with it, you can think of it like a thousand way multiple choice test. So the ImageNet competition would have universities and corporate research labs submit source code to compete in it.

[00:04:37] And the way it works is you hold up a picture and you'd ask, what is this of? A human can do that relatively easily, but the challenge would be out of a thousand or so potential categories. How do you make algorithms that can identify? And what happened around that time was humans were no longer better than computers at identifying objects and images.

[00:05:01] The best algorithms beat a college educated human. And what I saw happen then was, seemed like the entire Silicon Valley zeitgeist was pointing at self-driving cars and drones and these very competitive markets, super ambitious problems to solve. I mean, even now as many years later, still not solve self-driving cars, still have edge cases, failures scenarios.

[00:05:22] We still have steering wheels on our cars. And what I wanted to do was find an industry that was not going to see that advancement in technology coming and then just provide an amazing user experience so that people would be able to to realize the benefits of the tech. And my wife was working at a moving company.

[00:05:39] So that was kind of the, the genesis where I had the tech background. She was working at a moving company. She was working in logistics. So if there was an issue and maybe a 12 foot truck was sent when you should have sent a 16 foot truck, it was her job to pick up the phone, figure out where do you get the larger truck from, and kind of manage all the downstream problems.

[00:05:59] And the more we learned about the space, the more I realized. It's not for lack of trying, it's just really difficult to provide an amazing experience because there's so many details you need to get right. In a typical home when you're moving, you may have 300 or so items, and all of them can be a little bit different.

[00:06:14] So how do you plan for a move? How do you make sure you have the right number of boxes on the truck? Seems like kind of simple, but if you bring. The wrong number. You have to come back the next day or go and pick up some more when you're supposed to be on the job. So it's just a very difficult problem to be able to solve, and that's why it seemed like ripe for computer vision to come along and help.

[00:06:35] Omer: Okay, great. So you, you, you, you've got the technology and you, you, you're seeing it, you know, applied in some of, you know, the, the kind of places you, you mentioned, but you, you're seeing also these problems, you know, in kind of more like day-to-day world with like moving companies and you're thinking, why, why do people have to go out?

[00:06:56] Why can't we use that same technology to solve this type of problem? What did you do to go and validate the idea? Did you, you know, I mean obviously you, you, you spoke to your wife and, and that's kind of, you know, one insider that you're getting some, some information from. Did you start to, you know, try and do the whole kind of thing, like line up interviews and go and talk to people running, moving companies?

[00:07:23] Zach: That's exactly what we did. I think one of the cool parts of the moving industry is it is super fragmented. I think there's 7,000 or so licensed movers in the US all the way, like large companies all the way down to like two guys in a pickup truck. And what we did was we first to validate, you could tell I'm an engineer, we wanted to see how big of a problem is this really.

[00:07:44] So I went to the Better Business Bureau and they have rankings of complaints that have been filed by industry, and we just pulled the data down. Turns out movers get complained about more than lawyers, more than diet supplements and more than airlines. So we figured, okay, there's, there's something here going on.

[00:08:00] And then we wanted to zoom in and see what exactly is it about moving companies that people complain about. And we looked at Yelp reviews, Google reviews, BBB complaints, and, did some rudimentary data analysis on the keywords that were being used, and what we found was people generally complain if you break their things, if you quote incorrectly, or if you don't show up on time.

[00:08:24] So we're a software company. We, we can't necessarily help with the showing up on time part because that's a physical, physical thing, but the prices being incorrect was something that we learned was really, really an issue. So then from that, we're able to go. Call multiple moving companies. And again, cool part about it being a fragmented space is I can get hung up on a bunch and that's okay because there's so many fish in the sea.

[00:08:47] So we did some cold calling. I randomly showed up to a couple that didn't go over too well usually just get escorted out of the building, but I. In the beginning days, you gotta do things that don't scale.

[00:08:57] Omer: So how, so how did you, how did you kind of, you know, you, you, you, we were chatting and you were saying, look, I'm, I'm an introvert, I'm an engineer.

[00:09:05] How, how hard was it to just walk into these places or start cold calling and, you know, once you get those first few rejections, did it kind of make it harder or, or did, did that just. You know, kind of smooth the, the, the kind of the, the wheels for you and, and make, you know, kind of just get you into get, get you into the flow of, of doing this.

[00:09:25] I'm, I'm just trying to understand like what was going through your head or how easy or, or hard it was to, to go and just talk to these people.

[00:09:34] Zach: Yeah, it gets easier, I would say it never got easy. I remember the feeling in the pit of my stomach. When you park the car, you're in the office, and even if they agreed to meet with you.

[00:09:43] And opened my phone for the 18th time looking for an excuse to not have to do it. So there's a certain amount of convincing myself that I had to do to more of like a mindset change where this is an impediment to progress if I want to see if this thing is gonna work. And I don't wanna just have like a fantasy in my head of one day running a company, I'm gonna have to get over this.

[00:10:02] And yeah, you have some embarrassing situations that come up, but I think if something's really important. You can convince yourself that it's worth going through the pain to get through it. And I think the, like in a lot of things in business and in life, I mean, you practice a bit, you get better and maybe you fudge one or two.

[00:10:21] But we would also work that into the system. So if you have a really important client and you really we want to do a pilot with them or you really value their feedback, put 'em fourth or fifth on the list so you can get some reps through the system. All learning is good learning, so just you can kind of be prioritizing along those things.

[00:10:38] So I would say it wasn't, it wasn't necessarily easy. I still don't think I'm the best person in the world at doing it, but the goal is to be good enough to get to the next level, not perfect.

[00:10:48] Omer: Yeah. Yeah. That's good attitude. Okay. So when you weren't being escorted outta the building and you managed to get these people's time, whether it was on the phone or in person, what was the reaction when you tell them that?

[00:11:03] We are gonna build this AI solution to do these, you know, virtual home surveys. Like, like were they excited about the idea?

[00:11:14] Zach: It was a very polarizing suggestion. There was about half of the folks that we would talk to would say, there, you can do this. You would completely revolutionize the industry. This is incredible.

[00:11:25] Like, can I buy it now? Sign me up. And then we didn't have it working yet, so we, we made a wait list for those folks. The other half though. Were skeptical and they would just say, yeah, right. If this could have been done, it would've been done by now. And that's where as an engineer and as a founder, you usually want to have some secret that you know that most of the market doesn't know.

[00:11:45] But you gotta remember that image net thing had happened. So I knew computers are better than people at identifying objects and images, but that's a big radical, like fundamental sea change. So it made sense that people didn't quite understand yet. So for those folks, we didn't say no forever. We just didn't continue following up with them.

[00:12:04] They maybe aren't an early adopter. Kept it in the CRM and put a note on it and if you're specific about the objections and if they mention, Hey, this is never gonna work, you can follow back up a year later when you have it working and say, Hey, would you like to try it out, here's a link free of charge.

[00:12:20] We'll happy to walk you through a demo, but yeah, it was very, very polarizing. People would immediately fall into one of those two categories.

[00:12:27] Omer: How long did it take to get to a point where you felt confident enough that there was an opportunity here, there was a startup there, potential and, and that, you know, you guys were gonna invest more time in, in building this product?

[00:12:43] Zach: Yeah. It took about maybe two, three months to convince myself that if the tech existed. That there would be a business there. We were also, for the conversations that went well, we were experimenting with different business models. So we even negotiated pricing and did these non-binding letters of intent.

[00:13:00] I mean, they're not really that they don't carry any legal weight, but they would help us articulate if such a thing were to exist is a hundred dollars an estimate, too much is $10. So we, we got it to the point where we had a, a list of clients that had basically agreed to pay reasonable rate per survey and do it assuming you're able to do X, Y, Z, A, B, C.

[00:13:22] So that took a couple months. The next question that came after that though is, is it possible to do what we were signing up to do and what these people are expecting? And that that part took a couple years to get really, really resolved.

[00:13:34] Omer: One question about the letter of intent. How did you frame that with these, these customers?

[00:13:40] Like how did you get them to sign something?

[00:13:43] Zach: So we did it as a wait list. We realized. We're not gonna have infinite capacity. I mean, Gmail took, what, 10 years to get out of beta. So we figured, hey, when this thing does exist, I'm not just going to open the floodgates to everybody. I'd like to take one client on, then three more than five more.

[00:14:00] So we pitched it as expressing interest in reserving a slot in there. So there was no cash, no upfront ask immediately, but it was, if you're interested and you'd like to be on this list. Would you be open to like monthly check-in calls and here's what the pricing would be. But again, you don't have to pay anything till it actually exists.

[00:14:20] And what we found then is you get the right kind of customer when you do that. You get someone who sees the potential of the technology, who wants to be first to market and everyone else. You'll need to build out more features, add more bells and whistles down the road. But you want your early adopters to be unapologetic supporters of what you're doing, and that process kind of helped filter those things out.

[00:14:44] If they started nitpicking about SLAs and uptime and can you translate into this language, they're all reasonable concerns. I. But just not for customer number one. So we would note the concern, park it, and then go look for somebody else.

[00:14:55] Omer: And, and did you actually get them to sign a piece of paper or, or e-signature or something?

[00:14:59] Zach: Yeah, we actually did Adobe fill and sign. Again, it was it was one page and it said, I think it was something to the effect of like, if Yembo had an AI powered virtual inspection solution that could identify objects from a customer's video, then we would be interested in paying. 10, 20, 50, a hundred, whatever dollars per survey for it.

[00:15:18] It was like very, very short and brief. But I think the point was to be able to articulate the value enough and then also you're not really negotiating. So you can be a little bit more abstract with it. So we'd say, would you pay a thousand dollars a survey? They'd say never. You kind of like learn how they're thinking about it, the expected ROI that they want to get out of it.

[00:15:39] And that really helped when we're building the product to understand. What is meaningful? What moves the needle for my customer? Because we don't wanna just extract a bunch of cash from our clients. We wanna help them go grow, win more business, and then participate in some of that upside.

[00:15:52] Omer: Yeah. I, I think that was a really smart thing you guys did.

[00:15:55] There's, when you, when you do these types of interviews and people say, that sounds great. It's tempting just to stop there and say they love it. And when we come back with the product, they're gonna pay. But I think to get to that last mile and actually have them sign something, even though as you say, you know, not legally binding or anything, there's a, there's another level of commitment or, or a data point that you're getting that they, that they're interested enough to be able to, you know, willing to do that.

[00:16:30] Right. So I think that was really smart. The other thing you said was it took a couple of years, or a few years for actually for us to actually build. The product that, you know, we wanted to wa was, was that, was there something in between, like, did you, did you come back with a first version or an MVP with these guys?

[00:16:52] How long was it between the time that these customers or potential customers were. Signing these Lois to the point where you came back and put something in front of them that they could start trying?

[00:17:02] Zach: Sure. Yeah. It, it was about a year or so from Lois signed to, here's a login probably about two or three years to the point where it was enough to be a viable business.

[00:17:14] And this, I think, was really key. And for any listeners who are thinking about applying AI or machine learning to your end-to-end products, I think this was a really key finding. If you expect the AI to be perfect, there will always be scenarios that doesn't quite work In AI is fundamentally probability based.

[00:17:33] Even a human right? If I, if I ask you like, Hey, read the text from the spine, what book is this? People aren't going to be perfect at reading it. So if you make your use case so high stakes that it needs to be perfect, you're going to have implementation issues. So what we did was we, we understood the customer's problem really well.

[00:17:53] They spend a lot of time burning fuel, sitting in traffic, wear and tear in the vehicle. It's not a particularly awesome customer experience to have a handwritten list of items. And then sometimes the prices change. You don't quite know why. So we had an intellectually honest value proposition, which was still intact if the AI wasn't perfect.

[00:18:15] So what I mean when I say that is typical home may have two, 300 different items. If our AI could only detect five or 10. We were still saving time and we were giving pictures of the actual items that were there in the, in the move. So the mover may have to go through and do some review and spend some time.

[00:18:34] It's not completely automatic, but I. They are providing better documentation, they're increasing their win rate. They can go into geographies, they don't physically have boots on the ground in, and there were enough reasons that someone would want to do that, that it was okay that the AI wasn't perfect.

[00:18:50] And that gave us a reason to exist another day. And then over time we were able to say, Hey, this product that we put together, the AI can detect 10 items, then 50, then 60, then a hundred, and we kind of gradually worked our way up. But we were really careful not to set the wrong expectation that it was going to be a completely driverless self-driving car, no steering wheel, and you can just go anywhere in the world by the click of a button on day one, because that would've set us up for failure.

[00:19:16] Omer: How did you guys, fund the business for the first few years.

[00:19:19] Zach: We did an early seed round with some angel investors and kind of bootstrapped that way. AI is really expensive. Just the compute is, is is a lot. But we tried to not get too hung up on raising gobs of cash and doing a bunch of crazy things.

[00:19:37] So we did make sure that, like the first dollar we took in was actually revenue. And then from there, the the investment was always to accelerate and to develop the product further and to, to make the AI better. And I think that was really key where we never really got too detached from reality that we wanted to make sure that the product that we're offering is valuable to our customers so that we see it as like an accelerator, but you have to already be on the right track and heading in the right direction.

[00:20:03] But we didn't wanna just like, take venture dollars and then go figure out what to do next 'cause that's a, that's a very inefficient way to operate.

[00:20:09] Omer: Yeah. Yeah. Okay. So you go back to those customers about a year later, give them a login, and then what happened? What was you, you'd, you'd, you'd been talking to them about this vision of how AI and this product could make their lives.

[00:20:31] Better. And, you know, some felt it was revolutionized the industry. What was the reality of that first version of the product?

[00:20:38] Zach: It was, it was bumpy, it was rough. So I think our AI could detect maybe 10 or so items really well. We had no full-time product designers. No one with psych backgrounds with the user interface was, it looked like it was built by a backend engineer me.

[00:20:53] So we had we had some churn problems. We had bad expectations where people were expecting it to see behind closed doors, work in the dark, just stuff that is like literally never going to happen. But like the promise was there and the right group of customers got value. So some people came and then left, but other people came and then expanded.

[00:21:15] And what we found was certain clients were able to expand the geographies they operated in. It's traditionally very expensive to open a new satellite office as a mover. You have to rent a new warehouse, trucks, crew, all these kinds of things. But with Yembo, you can buy Google AdWords in a new geography and quote the jobs and then not really actually drive out there unless you win it.

[00:21:36] So what we saw is people were able to decrease the cost of expanding their business. People will be able to service leads after hours. If someone's working nine to five, they may not be home to walk around and answer the door, but you could text them a link. They could do it at their own convenience.

[00:21:52] So what we found was the initial archetype of a customer that did well in the early days had one of those initial pain points. And then as the technology got better, as we added on more features, we were able to broaden the applicability. But what we found was that we needed that initial group where they felt some kind of pain point that our product at that point in time could serve.

[00:22:17] We don't like selling technology that's going to be valuable based on future performance. It's like if you sign in and you get a login, you should be able to do something today that's valuable and wasn't the easiest, but we were able to find that group of people then expand from there.

[00:22:31] Omer: Yeah, I, I think that.

[00:22:33] It's not just finding people with the pain, but like you guys were doing, it's also about finding those early adopters, people who are, who are more motivated or, or just kind of more inclined. To use this type of technology or to try things versus the people who have the pain. But the minute, you know, they type in their password wrong and can't log in, it's like they're game over.

[00:23:00] Right. They're not interested anymore. That's not your first ideal customer initially anyway.

[00:23:05] Zach: Yeah, and I think, and what if you don't, if you legitimately don't have capacity to serve everyone, then you can ask qualifying questions. And we've had a couple clients that. We're asking to be in the earlier cohort 'cause who, who, it's just human nature, right? Who would volunteer to be second? Right? Everyone wants to be first, but there were some folks who just mentioned, Hey, I don't think you're gonna get what you want or what you're expecting to get. If you go live with us on this day, I think you should go into the next group.

[00:23:29] And sometimes people bristle, but at the end of the day, you gotta, you gotta do what makes sense. And I just didn't want people to sign up with expectations that I know I couldn't live up to, because then you're gonna get a cancellation notice the next month when the renewal comes up. Right.

[00:23:44] Omer: So you said it was bumpy initially.

[00:23:48] Obviously over time the the AI got better, the product got better. That story about your customer's wife. We talked about when, when did that happen? Just tell us a story. I'm, I'm trying to figure out like, was that very early on or?

[00:24:06] Zach: Sure. Yeah. That, this was early on. So the, the Evo product for moving, someone scans a quick 22nd video.

[00:24:12] The AI summarizes it into some images and then shows what's there. So if you want to scan this room, you'd see tv, printer book cartons needed to pack it. And this particular room that we scanned was pretty busy. There's a lot going on, maybe 80 or so items, but the mover who was testing it out, his wife was standing in the room and she was kind of stretching like this a bit.

[00:24:35] And our AI accidentally tagged her as a surfboard.

[00:24:39] Omer: Surfboard.

[00:24:40] Zach: Which is not a kind of problem that a human would ever make, but if you're AI and you're just seeing lines and shapes and colors, that person bent like this, you can kind of, you can kind of see why. But, those are the kinds of problems that that I think a lot of AI companies face, where from a technical standpoint, it's not really like a different kind of failure than calling a sofa a loveseat, but from a sociology standpoint, it's dramatically different.

[00:25:04] And those are the kinds of things that we, I, we had to go back and actually became a barrier to adoption is the AI would detect like 80 things correct. And everyone wants to talk about the one or two mistakes it would make. Also, we found laundry baskets and bedrooms were often being called barbecue grills because usually barbecue grills are covered.

[00:25:21] You just see fabric over some contours but like you and I as humans know, you don't really keep a barbecue grill like in a heap on your bed in your bedroom. So we had to build out in the early days, our AI team called it Ugly hacks. I called it our common sense engine and we'd say things like, if you see refrigerators in the bathroom, they're probably white panel doors.

[00:25:42] Just call it a door, don't call it a refrigerator. And we had all these like ugly hacks just to make the AI not make stupid mistakes because those are the kinds of perceptions like you're not. You're not associating your brand with trust if if you make like glaringly obvious mistakes. So it was kind of like he brought it up in jest.

[00:25:57] It was funny. Took him maybe like two years to stop bringing it up as a joke. But it did actually make us reevaluate how we were being perceived and the, the product did change as a result of that.

[00:26:08] Omer: Okay. Obvious. So obviously you're, you're improving the AI technology and you've got these. Wonderful hacks in place to, you know, help you make it through to, you know, the next level, you know, next kind of wave of improvements that you're, you're gonna roll out.

[00:26:22] What did you do to manage expectations with customers when they're excited about the technology? And you said, you know, it was pretty accurate, but they were picking up on the 1, 2, 3 things that. It didn't recognize or do a great job with. So how did you manage that situation with, with customers?

[00:26:46] Zach: It was not a one and done thing.

[00:26:47] It's what you do every day. So in our sales decks, we made sure that we pitched it as something that saves time, but does not eliminate a human. We, put in a lot of energy and effort, like I don't feel great waking up in the morning building Terminator technology. So we made a point to say, Hey, these are commission based salespeople.

[00:27:04] You can close more jobs. Focus on growing your business and being intellectually honest about this. Actually, if people say that's not gonna happen, then like, ask them why, figure out why, and then come up with proof points and show that it's able to. To improve the top and the bottom line. So I mean, people do always bring up individual mistakes here and there, but I think it's, it's machine learning, it's probability based.

[00:27:27] People understand that but by having that core value prop around, okay, maybe we did call a sofa, let's seat, or a wife, a kayak or a surfboard or something, but. You got a quote out and the person came to your website at 7:00 PM when no one was in the office, would you have preferred that they just shopped around somewhere else and you would've lost the business?

[00:27:48] So we kind of like made sure that people were understanding it the right way, but then also every time they were right, we'd make sure that we'd catalog it. We know we had telemetry. If the AI was getting corrected, even if you never said anything, we would still know because like any AI product, you always want to be iterating and always want to be improving.

[00:28:05] Omer: I wanna talk a little bit about sales. Getting to that first million in ARR, you know, both, both you and your co-founder, Sid, are engineers. And many founders in that situation would would try to get a sales person or, or some kind of, you know, growth market or whatever on board as quickly as possible so they didn't have to talk to customers or try to sell anything.

[00:28:32] You guys were pretty deliberate and you, you, you decided that you were gonna do the selling even though you had no experience. Why, why did you do that and, and kind of what was the experience for you?

[00:28:45] Zach: Sure. I think it made sense for the time and place we're at today. We have sales folks. They're better than I am at it, and I don't necessarily miss those days.

[00:28:52] However, in the early days when you're selling nascent technology. The market doesn't quite understand it yet. You don't even quite a hundred percent know. Is it gonna be perfect? Is it gonna work? Is it, is it gonna be viable or not? I didn't want to have another layer in between on hearing that feedback and getting those objections.

[00:29:10] And I think it goes back to those early days when I was telling you, when we were looking at the better business reviews, we were talking to prospects is I got told directly this is important to me. I wouldn't pay for that. And we decided until we got to a million in ARR, I didn't want to be trying to outsource it 'cause I don't know what works or not.

[00:29:26] If I bring someone on and they can't close any deals and they come back and say the product doesn't work, I didn't have enough history with it to really understand what's right, what's wrong. So I don't think I was the best salesperson. I would pretty much cave when any objections came up and maybe despite ourselves the product was good enough, we were able to get to that milestone.

[00:29:46] But when you're still setting it up, you want that direct feedback line. And if I build something that I think is going to be amazing, and it's not. I don't want it to take two weeks to filter back to me. I want to just be told directly from the customer what's working and what's not. So I think in that kind of environment, it was pretty clear when it was time to hand it off.

[00:30:06] I mean, a million is kind of like an arbitrary number, but what we were what we saw was the process was starting to become repeatable. It was, we started to look at things like cycle time. How long does it take to close a deal? What are the common questions that come up? What are common payment or pricing objections, all these kinds of things.

[00:30:23] I wasn't really learning anything new by doing it anymore, so that was a good sign that it was probably time to grow up. Let somebody else take that over and then hand it off.

[00:30:32] Omer: How were you generating leads? How were you finding these customers?

[00:30:36] Zach: So the moving market is very networked, so we found trade shows worked really well for us.

[00:30:41] If I have an office in San Francisco, you have an office in New York. We're not really competitors. Like if someone's moving across country, I may hire, I may have my crew go pack up the home on the origin, and then your crew goes on the destination. So what we found was. Going to these trade shows where all these folks are at, and then referrals around.

[00:31:00] If you have a happy customer, you're able to improve their top and bottom lines. Finding out who do they tend to work with and kind of working that angle has worked out pretty well for us. But I think it just comes down to finding, I. Who's hungry for the value you're providing and how do you bring it to them?

[00:31:15] And what we found for us is movers are all over the world, but when you have a conference, everyone's in one room. So that worked out great for us.

[00:31:22] Omer: Yeah. And it, it turns out that events and trade shows have turned out to be a great, great growth channel for your business. It's something you still do today.

[00:31:33] I think that. I'd love for you to explain like, how you were setting things up in a booth and, and helping people experience, you know, eYembo rather than just telling them about it. But I think it's also funny because, you know, in your book, which we'll we'll talk about in, in a, in a couple of minutes, the first thing you talk about is like.

[00:31:56] You know, having a booth at an event and it was like 900 bucks and it's like, do we really wanna spend $900 on a booth? Right. And that's the reality for many early stage startups, right? It's, that's a lot of money. But it, you know, it took, it turned out to be a, a good bet and you know, a great way to meet customers and, and generate sales.

[00:32:17] But yeah, just tell us about like, you know what, what was that Booth experience like? What were you exactly, what were you doing?

[00:32:22] Zach: Sure. So year one, you're absolutely right. Sponsorship was $900. I'm an engineer. I did the math. I said, I can sit at the bar and buy $9 beers for a hundred prospects, and it'll be the same as getting this booth.

[00:32:34] We didn't have the product yet, so we did that. Second year, the product was a bit more mature. The venue actually changed the rules and said, you can't just freeload in the bar area unless you have a pass. So we did get the booth there and what we were finding is people were generally interested but skeptical because a lot of these folks have been doing surveys inside people's homes for 20, 30, 40 years.

[00:32:56] So to come along and say, oh, I have AA that does it, people would kind of roll their eyes and say, yeah, right. So our booth was very simple. Again, I'd never designed a booth before. Probably wouldn't pay me to decorate a room or anything, but we had that idea in mind around people are gonna be interested but skeptical.

[00:33:12] So what we did is we flew out to, the event was in Florida, so I'm, I'm in San Diego, fly across the country, took an Uber xl, the big like minivan comes, picks you up. Went to TJ Maxx and just bought some furniture. We got a sofa chair, a loveseat, a lamp or a, a little nightstand, a lamp. And the booth was just putting the furniture there.

[00:33:33] And when we told people what we were doing, they would say, oh, really? And I'd just hand them a phone. I'd say, yeah, point it, point it at the furniture, and you'll see the results right away. And they'd go and they'd do that. And then like, their faces, it was cool. It was like, you're a magician, almost. Like their faces would light up and then they'd start objecting.

[00:33:49] Well, of course you, you pick that furniture. So I, I brought the receipt and I said, no. Look, this was purchased like 45 minutes ago when I was putting this demo together back in my office. I had no clue what I was gonna find. I just wanted to see what would fit in the Uber. And what we found is that like being able to do it live and be real shows that it's you're a credible person and that you're not pitching vaporware and that, that kind of experience.

[00:34:12] We've had a couple trade shows like that in the early days where, there was one time we even paid for it. We're in the black before we even came back home because we closed enough deals Just from, from that like on the spot convincing. But I just think it's being intellectually honest, being able to be willing to do it live.

[00:34:26] And it comes back to our engineering roots around, we're a technical company and we stand behind what we do and I. Is it the best way to get sleep the night before? Absolutely not. But it makes her a really convincing demo just to, to do it real and do it live.

[00:34:38] Omer: Yeah. That's great. Show me, don't tell me.

[00:34:40] Right. That's exactly what you were doing. Love that. Okay. We should we should wrap up. Let's get into the lightning round. I've got seven quick fire questions for you. So whenever you're ready. Okay. What's one of the best pieces of business advice you've received?

[00:34:54] Zach: Just get started. You don't really know what you're gonna get told until you actually do it, so don't, don't convince yourself in your head.

[00:35:01] Just get started.

[00:35:02] Omer: What book would you recommend to our audience and why?

[00:35:04] Zach: I think Zero to One by Peter Thiel is a good fundamental exposition on startups and how to, how to change an industry and disrupt the world. I think any, any founder should read it if they haven't already.

[00:35:16] Omer: Great. And then we also gotta mention your book.

[00:35:18] You, you wrote the book called Grow Up Fast Lessons from an AI startup.

[00:35:23] Zach: I got, got one right here.

[00:35:25] Omer: Awesome. So people can, we'll, we'll, we'll talk about where, where people get that in a second. But how did you find the time to write a book?

[00:35:31] Zach: A lot of little, a lot of little things. I I was disciplined. I didn't binge, I was disciplined.

[00:35:36] I booked 90 minutes each morning before the workday, around six to seven 30 in the morning. I did three days a week because I figured five is too ambitious and it took about a year. But what I was finding was there was just so much happening in the AI space. I felt like I had things to share. I was repeating myself a lot to like new managers who were hiring things along those lines.

[00:35:56] So I wanted to take the time to package it up and, and share it with a wider audience.

[00:36:00] Omer: Yeah. Love it. What's one attribute or characteristic in your mind of a successful founder?

[00:36:07] Zach: I would say resilience is key, is that anyone can conquer the world on a good day, but to be told no, to have a setback and to be able to brush your ego aside and figure out what you're gonna do about it, I think that's what separates true founders from people who are just interested in startups.

[00:36:24] Omer: What's your favorite personal productivity tool or habit?

[00:36:27] Zach: This is gonna sound really low tech, but my to-do list is I email myself and I use inbox zero. I've tried every other tool on the planet, but it's just so hard to beat and I'm in my inbox all day anyway. Yeah.

[00:36:37] Omer: What's a new or crazy business idea you'd love to pursue if you had the time?

[00:36:41] Zach: I think there's a lot of non. Tech areas where AI can be impactful. I would love to open an art gallery powered by AI or one of these industries where they traditionally haven't been served. But you can take a new and interesting angle on it. Again, I think I got time to write a book. I don't think I have time to do that today, but maybe 10 years from now.

[00:37:00] We'll see.

[00:37:01] Omer: What's an interesting or fun fact about you that most people don't know?

[00:37:04] Zach: I spent five years living in Vermont and I grew up with two llamas. Wow. And I, it's one of those things you gotta show, don't tell. So I had to go back to my parents' house last Thanksgiving and scan some photos because nobody believed me.

[00:37:16] So I had to be able to have photos I could send out one.

[00:37:18] Omer: That's funny. And finally, what's one of your most important passions outside of your work?

[00:37:23] Zach: I've got a family. I got three kids all under the age of six, so don't have a ton of time left over after hanging out with them, but love the outdoors was just out in a cabin in the woods with them last week.

[00:37:33] And I just think spending time with with family hanging out, doing probably boring things that you wouldn't wanna talk about on a podcast, but meaningful and fulfilling things with with close friends and family.

[00:37:44] Omer: I gotta say, you're, you, you, you always struck me as a very chilled guy. For somebody working on a startup and having three kids under six, it's not like that's a lot of stress there.

[00:37:55] Zach: It's a, it's a learn, learn skill. Yeah. I think me freshman year in college was super not chilled and you just kinda learn that. You learn to trust your problem solving skills as you get older is that it's not like I've seen it all before, but I know how to handle it if something comes up.

[00:38:07] Omer: Yeah. Okay, great.

[00:38:09] So tha Zach, thank you so much for joining me. And also it's your birthday today, so happy birthday. I appreciate you making the time today. If people wanna learn more about Yabo, they can go to That's yembo[dot]ai. If people wanna check out your book, they can go to Or find it on Amazon, we'll include links in the show notes and if folks want to get in touch with you, what's the best way for them to do that?

[00:38:37] Zach: Probably the easiest is find me on LinkedIn, Zach Rattner, or you can type in Yembo. It's not hard to find me, but I'm on there almost as much as I'm in my inbox, so feel free to send me a note.

[00:38:46] Omer: Sounds good. Thanks man. I appreciate you making the time great conversation. Congratulations on everything you guys have accomplished so far.

[00:38:54] And I wish you and the team the, the best of success.

[00:38:57] Zach: Thank you so much, Omer. Omer. Happy to be here.

[00:38:59] Omer: Cheers.

Book Recommendation

The Show Notes